334 



THE AMERICAN NATURALIST. [Vol. XXXVIII. 



P. insignis the transition forms exhibit much more detailed 

 gradations by virtue of which it is often exceedingly difficult to 

 distinguish between the two forms of cell, since whether coter- 

 minous or parallel the variations in thickness change in such a 

 way that the one type passes gradually into the other. When 

 these variations are viewed collectively and taken together with 

 the general fact that the thin-walled cells are a feature of the 

 higher types of organization, we may reasonably conclude that 

 the^thin-walled cells have been derived from the thick-walled 

 through a process of arrested development. The cause of such 

 alterations is to be sought for, and it will doubtless be found in 

 connection with another component of the ray. 



The terminal walls of the ray cells present three typical vari- 



ations : (I) thin-walled and entire; (2) thin-walled and locally 

 thickened, and (3) thick-walled and coarsely pitted (see table, 

 D of anatomical data). The first feature is a characteristic 

 of 52.6 ^ of all the genera, inclusive of Ginkgo, from Agathis 

 to Sequoia, while it also appears in Cupressus and Abies in part 

 as exceptional, and in the genus Pinus to the extent of 85.3%- 

 The wall presents no secondary growth in thickness, either 

 locally or generally. In the majority of cases it crosses the line 

 of the principal cell axis either at right angles or diagonally, 



